Gene Curation at the Rat Genome Database

1
Gene Curation at the Rat Genome Database
Stan Laulederkind, Mary Shimoyama, Andrew Patzer,
Victoria Petri, Jennifer Smith, Rajni Nigam,
George Kowalski, Dawei Li, Jeff De Pons, Melinda
Dwinell, Simon Twigger, Howard Jacob Rat Genome
Database, Bioinformatics Program, Medical College
of Wisconsin
Disease Portal Project
Abstract
Disease-gene Databases
During the past year both informatic and manual
gene curation at RGD have gone through efficiency
improvements. The implementation and enhancement
of multiple informatic pipelines from external
sources has expanded and improved data at RGD.
Manual gene curation, once dominated by free text
annotations and spreadsheet entry of ontology
terms, has been converted to the exclusive use of
ontologies and the use of proprietary curation
software. The software allows easy and
simultaneous entry of curation from four separate
hierarchical ontologies gene ontology (GO),
disease ontology (MeSH), pathway ontology (RGD
proprietary ontology), and mammalian phenotype
ontology (MP from MGI). The improvement in
manual curation has resulted in an increase of 2
to 3 times the rate of annotations made compared
to using the previous methods. Most of the
current manual gene curation is focused on
specific data sets disease portals, OMIM genes,
specific pathways, and the reference genome. An
obesity/metabolic syndrome portal has recently
been deployed, joining the previously released
neurological and cardiovascular disease portals.
The process of curating genes for the disease
portals begins with the generation of a
prioritized gene list. Various human disease
databases (GeneCards, Genetic Association
Database, and Genatlas) are queried with
descriptive terms associated with the targeted
disease category. Beginning with the literature
for the most widely cited genes, we curate across
the four ontologies simultaneously. We curate rat
genes with gene ontology, rat and human genes
with phenotype ontology, and rat, mouse, and
human genes with disease and pathway ontologies.
The process is both thorough and efficient. To
fill in gene ontology information for human, data
is downloaded from GOA (http//www.ebi.ac.uk/GOA/)
. To fill in gene ontology and mammalian
phenotype ontology information for mouse, data is
downloaded from MGI (http//www.informatics.jax.or
g/).
Gene Lists
Reference Genome Project
OMIM Project
Literature
Pathway Project
Number of manual annotations doubled in one
year
Disease Ontology (MeSH)
2-3x increase in curation rate
Curation Software Tool
22,000 manual annotations
Disease Portal at http//rgd.mcw.edu
Orthology Pipeline (MGI)
Annotation Pipeline (GOA)
Gene Records Pipeline (Entrez Gene)
Nomenclature Pipeline (MGI, HGNC)
Protein ID Pipeline (IPI)